Skip to content

Quick Start

This quick start guide is the abridged version to get you running quickly. Please also see the more detailed documentation for each step in the workflow:

Create your first recognition job

Run the Textum executable from the extracted folder (see Installation).

  1. Choose job type

  2. Printed Text - Kosmos-2.5 for printed documents

  3. Handwritten Text - TrOCR for handwritten documents

  4. Select input

  5. Single image (PNG, JPG, TIFF, etc.)

  6. PDF document (processes all pages)

  7. Folder of images

  8. Configure settings

  9. Job name

  10. Language (German, English, Dutch), used for detecting possibly unknown words / OCR mistakes
  11. Double page splitting (if needed)

  12. Rotation (if needed)

  13. Start processing - Monitor job progress in the Job Control view

Review Results

  1. Go to the Job Control view
  2. Click the Review button on a completed job, the Review Interface will open
  3. Switch pages with arrow buttons or dropdown in the top-left
  4. Save your progress before switching pages
  5. Mark pages as Done (progress is indicated in the page dropdown selector)

Export

  1. From the Review Interface or the Job Control View, click the Export button. Select if only current page or all pages should be exported
  2. Choose format: Plain Text, Word Document (DOCX), or PDF
  3. If you selected a language during job creation, unknown words were highlighted using word lists. Per default, highlight markers (==this text will be highlighted==) are not exported.
  4. Select destination
  5. Export complete

▶️ Next Steps

Again, if anything is unclear, please see the more detailed pages: